NVIDIA (NVDA-US) delivered its strongest weekly stock performance in over a year, fueled by two primary factors: high-profile backing for its artificial intelligence (AI) chips and sustained large-scale investments in AI infrastructure by major cloud operators, underscoring strong market demand.

NVIDIA closed Friday (7th) up 2.3%, marking a cumulative gain of 11.6% for the week—the best weekly performance since May 16, 2025, when it surged 16.1%. According to Dow Jones Market Data, NVIDIA's market capitalization increased by $562 billion this week, setting a new record for the largest single-week market cap growth in the company's history.

NVIDIA's upcoming Vera Rubin chip architecture received strong praise this Tuesday from SpaceX (SPCX-US) CEO Elon Musk. During SpaceX's first earnings call, Musk stated that the company will "decide to use only NVIDIA" chips going forward, adding, "We believe this is the best AI computer, and we highly value our close collaboration and partnership with NVIDIA on multiple levels."

As SpaceX released its earnings, Google (GOOGL-US), Amazon.com (AMZN-US), and Meta Platforms (META-US) also reported their results, each raising their capital expenditure forecasts for the year.

Kate Leaman, Chief Market Analyst at AvaTrade, noted that beyond Musk's high praise, "the bigger story is the scale of current spending by major cloud providers on AI infrastructure," with NVIDIA remaining the dominant supplier of graphics processing units (GPUs).

Willy Lee, Partner at investment firm Neostellar, said that although Google's negative free cash flow for the quarter raised investor concerns, Microsoft's (MSFT-US) "more prudent approach to capital expenditures" and rapidly growing cloud business have led investors to view this wave of AI spending more positively.

Leaman added that with NVIDIA scheduled to report earnings on August 26, amid extremely high market expectations, "some traders have clearly positioned themselves early." However, she cautioned that given already elevated expectations, the stock could still pull back.

Meanwhile, Musk stated that SpaceX plans to deploy nearly 10 gigawatts (GW) of computing capacity next year, and that there is a "consensus" between SpaceX and NVIDIA that the company will secure a "very large proportion" of NVIDIA's GPUs next year.

Stefan Slowinski, analyst at BNP Paribas, said Musk's comments raise a critical question: Could GPU supply once again become a "key bottleneck" for AI infrastructure deployment over the next 12 to 18 months? This presents both risks and opportunities.

Slowinski pointed out that computing service providers without strong strategic ties to NVIDIA may fall behind. Conversely, companies able to secure allocations of NVIDIA chips or expand their own custom chip development programs "could reap disproportionate benefits."

He believes that if NVIDIA faces supply shortages, CoreWeave (CRWV-US), Nebius Group (NBIS-US), and Oracle (ORCL-US) would be best positioned to obtain large volumes of NVIDIA chips. However, these firms also face the risk that NVIDIA might prioritize expanding its partnership with SpaceX.

He noted that SpaceX's emphasis on securing NVIDIA chip allocations "further strengthens the view that AI compute shortages are unlikely to ease before 2027."

Additionally, SpaceX indicated that memory output grows about 20% annually, while AI demand increases by as much as 200%. Slowinski said this supports the view that "supply remains severely behind demand." Therefore, even if the above figures are only directionally accurate, it implies that the current favorable pricing environment is unlikely to normalize in the short term.

FACT BOX

  • Source: PR Times
  • Category: Partnership
  • Organizations: SpaceX / Google / Amazon
  • Products / services: GPU